RECORDING SPEECH VIBRATIONS. 21 



investigation would be a matter of importance in reference to the action 

 of sound-receiving instruments such as talking-machine recorders, tele- 

 phone transmitters and the human ear. 



It is at present not my intention to pursue these investigations further. 

 What has already been done indicates that the formation of nodes (Chladni 

 figures), as in most vibrating plates, is a matter of no importance in the 

 stiff diaphragms of talking machines. The essential factor is the curve 

 of bending. This depends on the material and on the edge-fixation. In 

 my opinion the aim should be to obtain a rigid diaphragm with Hquid 

 fixation. The diaphragm would yield to the vibration pressure hke a 

 piston in a cylinder. The distortion of the vibration would depend on 

 the inertia and the friction. 



The sound boxes now made for gramophones and phonographs are 

 the results of experience without reasoning. According to the prevalent 

 view of their action (as far as any view at all is prevalent) the diaphragms 

 are regarded — without justification — as Chladni plates. The above 

 experiments indicate the ideal toward which the constructors of sound 

 boxes unconsciously strive. It comprises a stiff circle in the center 6, 

 figure 17, held in position l^y a ring, a, that acts as a spring and also 

 as an air-tight closure. The tube d has the diameter of the stiff center. 

 Over the ring a enough air space is left in the frame to allow free move- 

 ment. The recording or reproducing link c acts like a piston-rod attached 

 to the stiff center. It is essential that the stiff center b should have per- 

 fectly regular movement; therefore the spring ring a must be perfectly 

 homogeneous. This can not at present be accomplished without making 

 the diaphragm all in one piece. Consequently, in order to obtain springiness 

 in the outer portion the stiffness must be yielded in the center; the glass 

 or mica diaphragm thus bends at all points. A new model of box partly 

 recovers the central stiffness by additional central layers of mica. 



If we rule out any nodal action in the diaphragm, what is then the 

 cause of the gramophone and phonograph twangs? The two are different 

 from each other and from the natural sound. Let us put to one side 

 the fact that the small diaphragm of the talking machine is physically 

 incapable of creating such strong vibrations as those of a bass drum; we 

 will also disregard further questions of loudness and the modification of 

 the sound by the trumpet. 



The falsification of natural sounds by the talking-machine recorders 

 occurs, in my opinion, from a distortion of the waves by the bending of 

 the diaphragm and not to nodal vibrations. The sinusoid movement 

 imparted to the rod by the fork, as described above (p. 17), would, with 





